Downscaling of air pollutants in Europe using uEMEP_v6

The air quality downscaling model uEMEP and its combination with the EMEP MSC-W chemical transport model are used here to achieve high-resolution air quality modelling at street level in Europe. By using publicly available proxy data, this uEMEP-EMEP modelling system is applied to calculate annual mean NO2, PM2.5, PM10, and O-3 concentrations for all of Europe down to 100 m resolution and is validated against all available AIRBASE monitoring stations in Europe at 25 m resolution. Downscaling is carried out on annual mean concentrations, requiring special attention to non-linear processes, such as NO2 chemistry for which frequency distributions are applied to better represent the non-linear NO2 chemistry. The downscaling shows significant improvement in NO2 concentrations for which the spatial correlation has been doubled for most countries and bias reduced from -46 % to -18 % for all stations in Europe. The downscaling of PM2.5 and PM10 does not show improvement in spatial correlation but does reduce the overall bias in the European calculations from -21 % to -11 % and from -39 % to -30 % for PM2.5 and PM10, respectively. There is improved spatial correlation in most countries after downscaling of O-3 and a reduced positive bias of O-3 concentrations from +16 % to +11 %. Sensitivity tests in Norway show that improvements in the emission and emission proxy data used for the downscaling can significantly improve both the NO2 and PM results. The downscaling development opens the way for improved exposure estimates, improved assessment of emissions, and detailed calculations of source contributions to exceedances in a consistent way for all of Europe at high resolution.